The content distribution network or content delivery network (CDN) is one kind of new network framework, which can release the service content to the “edge” of the network nearest to the user, to make the user can obtain the nearest required content and improve the operating efficiency of the system. Along with the rapid development of the network, the CDN is used in the commercial system in the real life, such as internet protocol television (IPTV), global eyes and so on, which has a very good prospect. In these commercial systems, the CDN is generally operated coordinating with other modules, which mainly includes the stream media system and some corresponding terminal equipments. The CDN belongs to the service layer, and mainly used to schedule and process the related service; and the stream media system mainly takes charge of implementing the service.
The CDN is mainly composed of the CDN manager and the CDN node, wherein one CDN node is composed of several CDN agents (node controller) and other attached equipments. The CDN manager takes charge of the traffic load balance of the CDN and the distribution and scheduling of the content, and one CDN manager generally manages several CDN nodes. The CDN node is the service entity of the CDN, which is deployed at the edge of the network and provides services for the nearest user. The CDN agent is a CDN node controller, which takes charge of the load balance management of the node and the content management in the node, implementing the service control of the CDN node. At the same time, only one CDN agent controls the CDN node and this CDN agent is called node main controller, while other CDN agents are the node standby controllers.
The current implementing methods mainly use the main-standby controller changeover triggered by the manager as an original version, that is, the manager uses the heartbeat message detecting mechanism based on the wide area network. If the network is abnormal and the node controller does not respond for a long time, then the manager considers that the node controller is abnormal and set the current node main controller as standby, and then blindly searches other node controllers of which the heartbeat message is normal. In fact, in most cases, the status of the current node main controller is always normal; however, the communication problem of the wide area network makes the manager reject the current node main controller. Originally the current node main controller possesses the services capability, however, the heartbeat message to the manager is abnormal, then the manager navigates the services request of the requester to other node controllers through the service logic, which wastes the services capability of the current node controller greatly. Furthermore, if the wide area network is in a good and bad status for a long time, then the manager will certainly detect and switch all the controllers in the present node frequently, which causes the switch vibration and consumes the performance of the manager and the bandwidth of the network.